Preparation of GrapheneSelected Topics in Physics: Physics of Nanoscale Carbon

Nils KraneNils.Krane@fu-berlin.de

Today there are several methods for the preparation of graphene. In the following some of these meth-

ods will be presented and discussed. They will be compared using specific requirements and allocated

to different purposes. The requirements are e.g. quality and size of the flakes or controllability of the

resulting coating.

1 Introduction

Graphene is a very special material, since ithas the advantage of being both conductingand transparent. The transparency of a mate-rial normally depends on its electronic prop-erties and requires a band gap. Under nor-mal conditions transparency and conductiv-ity exclude each other, except for a few com-pounds like indium tin oxide (ITO). How-ever, in contrast to ITO, graphene is also flex-ible and capable of withstanding high stress.Therefore it is very attractive for the applica-tion of flexible electronic devices, e.g. touchscreens [1]. Accordingly, there are a lot of ef-forts in order to prepare graphene easyly withthe required properties.

The methods described in this review areevaluated on the bases of different require-ments: On the purity of the graphene, whichis defined by the lack of intrinsic defects,(Quality) as well as on the size of the obtainedflakes or layers (Size). Another aspect is theamount of graphene which can be producedsimultaneously (Amount) or the complexitysuch as the requirement of labour or the needfor specially designed machines (Complex.).One last attribute is the controllability of themethod in order to achieve reproducible re-sults (Control.).

This review is organized as follows: In Sec.2 the exfoliation methods are presented anddiscussed. Following that another type ofpreparation, the growth of graphene, will be

introduced in Sec. 3. Concluding, the differ-ent methods will be compared in Sec. 4.

2 Exfoliation

Basically there are two different approachesto preparing graphene. On the one handgraphene can be detached from an alreadyexisting graphite crystal, the so-called ex-foliation methods, on the other hand thegraphene layer can be grown directly on asubstrate surface. The first reported prepara-tion of graphene was by Novoselov and Gaimin 2004 [2] by exfoliation using a simple ad-hesive tape.

2.1 The “Scotch Tape Method”

In this micromechanical exfoliation method,graphene is detached from a graphite crys-tal using adhesive tape. After peeling itoff the graphite, multiple-layer graphene re-mains on the tape. By repeated peeling themultiple-layer graphene is cleaved into vari-ous flakes of few-layer graphene. Afterwardsthe tape is attached to the substrate and theglue solved, e.g. by acetone, in order to de-tach the tape. Finally one last peeling with anunused tape is performed.

The obtained flakes differ considerably insize and thickness, where the sizes rangefrom nanometers to several tens of microme-ters for single-layer graphene, depending onthe preparation of the used wafer. Single-

1

layer graphene has a absorption rate of 2%,nevertheless it is possible to see it under alight microscope on SiO2/Si, due to interfer-ence effects [3]. However, it is difficult toobtain larger amounts of graphene by thismethod, not even taking into account thelack of controllability. The complexity ofthis method is basically low, nevertheless thegraphene flakes need to be found on the sub-strate surface, which is labour intensive. Thequality of the prepared graphene is very highwith almost no defects.

2.2 Dispersion of Graphite

Graphene can be prepared in liquid-phase.This allows upscaling the production, in or-der to obtain a much higher amount ofgraphene. The easiest method would beto disperse the graphite in an organic sol-vent with nearly the same surface energy asgraphite [4]. Thereby, the energy barrier is re-duced, which has to be overcome in order todetach a graphene layer from the crystal. Thesolution is then sonicated in an ultrasoundbath for several hundreds hours or a voltageis applied [5]. After the dispersion, the solu-tion has to be centrifuged in order to disposeof the thicker flakes.

The quality of the obtained grapheneflakes is very high in accordance with the mi-cromechanical exfoliation. Its size howeveris still very small, neither is the controllabil-ity given. On the other hand, the complex-ity is very low, and as mentioned above thismethod allows preparing large amounts ofgraphene.

2.3 Graphite Oxide Exfoliation

The principle of liquid-phase exfoliation canalso be used to exfoliate graphite oxide. Dueto several functional groups like epoxide orhyroxyl, graphene oxide is hydrophilic andcan be solved in water by sonication or stir-ring. Thereby the layers become negativelycharged and thus a recombination is inhib-ited by the electrical repulsion. After cen-trifugation the graphene oxide has to be re-

Figure 1: (a) Solution of graphene in liquid-phase. The flasks contain solutions after cen-trifugation at different frequencies [4]. (b)Scheme of the exfoliation of graphite ox-ide. The graphite gets oxidized and solvedin water. Afterwards it gets reduced tographene [6].

duced to regular graphene by thermal orchemical methods. It is hardly possible todispose of all the oxygen. In fact, an atomicC/O ratio of about 10 still remains [6].

The performance of this method is verysimilar to liquid-phase exfoliation of pristinegraphene. Only the complexity is higher,since graphite oxide has to be producedfirst, wich requires the use of several chem-icals. Also the obtained graphene oxidehas to be reduced afterwards, using thermaltreatments or chemicals again [7]. The re-duced graphene oxide is of very bad qualitycompared to pristine graphene, neverthelessgraphene oxide could be the desired prod-uct. Graphene oxide modified with Ca andMg ions is capable of forming very tensilegraphene oxide paper, as the ions are cross-linkers between the functional groups of thegraphene flakes [8].

2.4 Substrate Preparation

There are different methods for substratepreparation in order to use the dispersedgraphene in a non liquid-phase. By vacuum

2

filtration the solution is sucked through amembrane using a vacuum pump. As a resultthe graphene flakes end up as filtration cakeof graphene paper.

The deposition of graphene on a surfacecan be done by simple drop-casting where adrop of the solution is placed on top of thesubstrate. After the solvents have evaporated,the graphene flakes remain on the surface. Inorder to achive a more homogeneous coat-ing the sample can be rotated using the spin-coating method in order to disperse the so-lution with the help of the centrifugal force.With spray-coating, the solution ist sprayedonto the sample, which allows the prepara-tion of larger areas.

3 Growth on Surfaces

A totally different approach to obtaininggraphene is to grow it directly on a surface.Consequently the size of the obtained lay-ers are not dependent on the initial graphitecrystal. The growth can occur in two differentways. Either the carbon already exists in thesubstrate or it has to be added by chemicalvapour deposition (CVD).

3.1 Epitaxial Growth

Graphene can be prepared by simply heatingand cooling down an SiC crystal [9]. Gen-erally speaking single- or bi-layer grapheneforms on the Si face of the crystal, whereasfew-layer graphene grows on the C face [10].The results are highly dependent on the pa-rameters used, like temperatur, heating rate,or pressure. In fact, if temperatures and pres-sure are too high the growth of nanotubes in-stead of graphene can occur. The graphitiza-tion of SiC was discovered in 1955, but it wasregarded as unwelcome side effect instead ofa method of preparing graphene [11].

The Ni(111) surface has a lattice structurevery similar to the one of graphene, with amissmatch of the lattice constant at about1.3% [11]. Thus by use of the nickel diffu-sion method a thin Ni layer is evaporated

onto a SiC crystal. Upon heating the car-bon diffuses through the Ni layer and formsa graphene or graphite layer on the surface,depending on the heating rate. The thusproduced graphene is easier to detach fromthe surface than the graphene produced bythe growth on a simple SiC crystal withoutNi [11].

Figure 2: SEM image of graphene on cop-per foil. At several locations on the sur-face graphene islands form and grow to-gether [14].

The growth of graphene starts at severallocations on the crystal simultaneously andthese graphene islands grow together, asshown in Fig. 2). Therefore the graphene isnot perfectly homogeneous, due to defects orgrain boundaries. Its quality therefore is notas good as that of exfoliated graphene, exceptthe graphene would be grown on a perfectsingle crystal. However, the size of the ho-mogeneous graphene layer is limited by thesize of the crystal used. The possibility to pro-duce large amounts of graphene by epitaxialgrowth is not as good as by liquid-phase exfo-liation, though the controllability to gain re-producible results is given. Also the complex-ity of these methods is comparatively low.

3.2 Chemical Vapour Deposition

Chemical vapour deposition is a well knownprocess in which a substrate is exposed togaseous compounds. These compounds de-compose on the surface in order to grow athin film, whereas the by-products evaporate.

3

Figure 3: (a)Scheme of preparation of graphene by CVD and transfer via polymer support.The carbon solves into the Ni during the CVD and forms graphene on the surface after cool-ing. With a polymer support the graphene can be stamped onto another substrate, after etch-ing of the Ni layer. Patterning of the Ni layer allows a control of the shape of the graphene [12].(b) Roll-to-roll process of graphene films grown on copper foils and transferred on a targetsubstrate [1].

There are a lot of different ways to achievethis, e.g. by heating the sample with a fil-ament or with plasma. Graphene can begrown by exposing of a Ni film to a gas mix-ture of H2, CH4 and Ar at about 1000°C [12].The methane decomposes on the surface, sothat the hydrogene evaporates. The carbondiffuses into the Ni. After cooling down in anAr atomosphere, a graphene layer grows onthe surface, a process similar to the Ni diffu-sion method. Hence, the average number oflayers depends on the Ni thickness and canbe controlled in this way. Furthermore, theshape of the graphene can also be controlledby patterning of the Ni layer.

These graphene layers can be transferedvia polymer support, which will be attachedonto the top of the graphene. After etchingthe Ni, the graphene can be stamped ontothe required substrate and the polymer sup-port gets peeled off or etched away. Usingthis method several layers of graphene canbe stamped onto each other in order to de-crease the resistance. Due to rotation rel-

atively to the other layers, the turbostraticgraphite does not have the Bernal stackingand consequently the single graphene lay-ers hardly change their electronic properties,since they interact marginally with the otherlayers [1].

Using copper instead of nickel as grow-ing substrate results in single-layer graphenewith less than 5% of few-layer graphene,which do not grow larger with time [13]. Thisbehavior is supposed to be caused due to thelow solubility of carbon in Cu. For this reasonBae and coworkers developed a roll-to-rollproduction of 30-inch graphene [1]. UsingCVD, a 30-inch graphene layer was grown ona copper foil and then transfered onto a PETfilm by a roll-to-roll process. CVD also allowsa doping of the graphene, e.g. with HNO3, inorder to decrease the resistance. Bae and col-leagues stacked four doped layer of grapheneonto a PET film and thus produced a fullyfunctional touch-screen panel. It has about90% optical transmission and about 30Ω persquare resistance, which is superior to ITO.

4

Figure 4: Assembled graphene/PET touchpanel shows high mechanical flexibility [1].

The mechanical performance test showed amuch higher withstanding of graphene com-pared to ITO. In fact, the resiliance was notlimited by the graphene itself, but by the at-tached silver electrodes.

In conclusion the author states that the op-tical and electrical performance of grapheneprepared by CVD is very high, but the pu-rity, which would be necessary for laboratoryresearch, is not given. On the other hand,the thus produced graphene layers can bevery large and are easily obtained in largeamounts. The complexity is rather low, sinceCVD is a well-known method and often usedin industry. Therefore there is no need to de-velop new machines or techniques. Further-more, perfect control of the results is given aswell as transportability.

4 Summary

An overview of the different methods andtheir performances is given in Tab. 1. In sum-mary, the exfoliation methods have the ad-vantage of providing graphene of very highquality and purity, and, due to the low com-plexity, they are perfect for laboratory re-search. The size of the obtained flakes, how-ever, as well as the controllability are too poorfor industrial production.

Method Qu

alit

y

Size

Am

ou

nt

Co

mp

lex.

Co

ntr

ol.

Adhesive Tape X × × (X) ×Liquid phase X × X X ×

Graphite oxide – × X × ×Epi. growth × (X) × X X

CVD × X X X X

Table 1: Overview of the performances of thedifferent methods.

On the other hand, the growth of grapheneon surfaces allows a more or less unlimitedsize of the graphene layers and a high con-trollability, which makes these methods ap-plicable for industrial production. The purity,however, is not very high, which makes thesemethods unsuitable for laboratory researchof graphene. Since CVD is an already usedmethod in industry, the epitaxial growth ofgraphene is probably a dead end technique.

References[1] Bae, S, et al.; Nature Nanotech. 5, 574-578 (2010)

[2] Novoselov, KS, et al.; Science 306, 666-669(2004)

[3] Casiraghi C, et al.; Nano Letters 7, 2711-2717(2007)

[4] Lotya M, et al.; ACS Nano 4, 3155-3162 (2010)

[5] Su CY, et al.; ACS Nano 5, 2332-2339 (2011)

[6] Park S, Ruoff RS; Nature Nanotech. 4, 217-224(2009)

[7] Tkachev SV, et al.; Inorganic Materials 47, 1-10(2011)

[8] Park S, et al.; ACS Nano 2, 572-578 (2008)

[9] Forbeaux I, et al.; Phys. Rev. B 58, 16396-16406(1998)

[10] Cambaz ZG, et al.; Carbon 46, 841-849 (2008)

[11] Enderlein, C; Dissertation: Graphene and itsInteraction with Di erent Substrates Studied byAngular-Resolved Photoemission Spectroscopy,Freie Universitaet Berlin (2010)

[12] Kim KS, et al.; Nature 457, 706-710 (2009)

[13] Xuesong L, et al.; Science 324, 1312-1314 (2009)

[14] Robertson AW, Warner JH, unpublished (2011)

5